GB2603445A - Identifying optimal weights to improve prediction accuracy in machine learning techniques - Google Patents

Identifying optimal weights to improve prediction accuracy in machine learning techniques Download PDF

Info

Publication number
GB2603445A
GB2603445A GB2207662.4A GB202207662A GB2603445A GB 2603445 A GB2603445 A GB 2603445A GB 202207662 A GB202207662 A GB 202207662A GB 2603445 A GB2603445 A GB 2603445A
Authority
GB
United Kingdom
Prior art keywords
student model
model
generating
training
candidate
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Withdrawn
Application number
GB2207662.4A
Other versions
GB202207662D0 (en
Inventor
Xu Jing
Er Han Si
George Barbee Steven
Ying Zhang Xue
Hui Yang Ji
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
International Business Machines Corp
Original Assignee
International Business Machines Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by International Business Machines Corp filed Critical International Business Machines Corp
Publication of GB202207662D0 publication Critical patent/GB202207662D0/en
Publication of GB2603445A publication Critical patent/GB2603445A/en
Withdrawn legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/02—Knowledge representation; Symbolic representation

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Computing Systems (AREA)
  • Artificial Intelligence (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Computational Linguistics (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Biophysics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

A computer-implemented method,system and computer program product for improving prediction accuracy in machine learning techniques. A teacher model is constructed: wherein the teacher model generates a weight for each data case.The current student model is then trained using training data and the weights generated by the teacher model. After training the current student model, the current student model generates state features: which are used by the teacher model to generate new weights. A candidate student model is then trained using training data and these new weights. A reward is generated by comparing the current student model with the candidate student model using training and testing data,which is used to update the teacher model if a stopping rule has not been satisfied. Upon a stopping rule being satisfied, the weights generated by the teacher model are deemed to be the "optimal"weights which are returned to the user.

Claims (25)

CLAIMS:
1. A computer-implemented method for improving prediction accuracy in machine learning techniques, the method comprising: constructing a teacher model, wherein said teacher model generates a weight for each data case; training a current student model using training data and weights generated by said teacher model; generating state features by said current student model; generating new weights by said teacher model using said state features; training a candidate student model using said training data and said new weights; generating a reward by comparing said current student model with said candidate student model using said training data and testing data to determine which is better at predicting an observed target; updating said teacher model with said reward in response to a stopping rule not being satisfied; and returning said new weights and said current student model to a user in response to said stopping rule being satisfied, wherein said returned student model provides a prediction of said observed target.
2. The method as recited in claim 1 further comprising: determining whether said candidate student model generates a better prediction of said observed target than said current student model based on how close the prediction is to said observed target.
3. The method as recited in claim 2 further comprising: updating said current student model with said candidate student model and updating current weights with said new weights in response to said candidate student model generating a better prediction of said observed target than said current student model.
4. The method as recited in claim 3 further comprising: generating new state features by said updated student model using said new weights; and generating a second set of new weights by said teacher model using said new state features.
5. The method as recited in claim 4 further comprising: training said candidate student model using said training data and said second set of new weights; and generating a reward by comparing said updated student model with said candidate student model using said training data and said testing data to determine which is better at predicting said observed target.
6. The method as recited in claim 2 further comprising: generating a second set of new weights by said updated teacher model using said state features in response to said candidate student model not generating a better prediction of said observed target than said current student model.
7. The method as recited in claim 6 further comprising: training said candidate student model using said training data and said second set of new weights; and generating a reward by comparing said current student model with said candidate student model using said training data and said testing data to determine which is better at predicting said observed target.
8. The method as recited in claim 1 , wherein said stopping rule comprises one or more of the following: reaching a specified number of trials, reaching a specified training time, converging of a prediction accuracy, and a user-initiated termination.
9. The method as recited in claim 1, wherein said teacher model comprises a neural network, wherein said student model comprises one of the following: a decision tree and a neural network.
10. A computer program product for improving prediction accuracy in machine learning techniques, the computer program product comprising a computer readable storage medium having program code embodied therewith, the program code comprising the programming instructions for: constructing a teacher model, wherein said teacher model generates a weight for each data case; training a current student model using training data and weights generated by said teacher model; generating state features by said current student model; generating new weights by said teacher model using said state features; training a candidate student model using said training data and said new weights; generating a reward by comparing said current student model with said candidate student model using said training data and testing data to determine which is better at predicting an observed target; updating said teacher model with said reward in response to a stopping rule not being satisfied; and returning said new weights and said current student model to a user in response to said stopping rule being satisfied, wherein said returned student model provides a prediction of said observed target.
11. The computer program product as recited in claim 10, wherein the program code further comprises the programming instructions for: determining whether said candidate student model generates a better prediction of said observed target than said current student model based on how close the prediction is to said observed target.
12. The computer program product as recited in claim 11, wherein the program code further comprises the programming instructions for: updating said current student model with said candidate student model and updating current weights with said new weights in response to said candidate student model generating a better prediction of said observed target than said current student model.
13. The computer program product as recited in claim 12, wherein the program code further comprises the programming instructions for: generating new state features by said updated student model using said new weights; and generating a second set of new weights by said teacher model using said new state features.
14. The computer program product as recited in claim 13, wherein the program code further comprises the programming instructions for: training said candidate student model using said training data and said second set of new weights; and generating a reward by comparing said updated student model with said candidate student model using said training data and said testing data to determine which is better at predicting said observed target.
15. The computer program product as recited in claim 11, wherein the program code further comprises the programming instructions for: generating a second set of new weights by said updated teacher model using said state features in response to said candidate student model not generating a better prediction of said observed target than said current student model.
16. The computer program product as recited in claim 15, wherein the program code further comprises the programming instructions for: training said candidate student model using said training data and said second set of new weights; and generating a reward by comparing said current student model with said candidate student model using said training data and said testing data to determine which is better at predicting said observed target.
17. The computer program product as recited in claim 10, wherein said stopping rule comprises one or more of the following: reaching a specified number of trials, reaching a specified training time, converging of a prediction accuracy, and a user-initiated termination.
18. A system, comprising: a memory for storing a computer program for improving prediction accuracy in machine learning techniques; and a processor connected to said memory, wherein said processor is configured to execute the program instructions of the computer program comprising: constructing a teacher model, wherein said teacher model generates a weight for each data case; training a current student model using training data and weights generated by said teacher model; generating state features by said current student model; generating new weights by said teacher model using said state features; training a candidate student model using said training data and said new weights; generating a reward by comparing said current student model with said candidate student model using said training data and testing data to determine which is better at predicting an observed target; updating said teacher model with said reward in response to a stopping rule not being satisfied; and returning said new weights and said current student model to a user in response to said stopping rule being satisfied, wherein said returned student model provides a prediction of said observed target.
19. The system as recited in claim 18, wherein the program instructions of the computer program further comprise: determining whether said candidate student model generates a better prediction of said observed target than said current student model based on how close the prediction is to said observed target.
20. The system as recited in claim 19, wherein the program instructions of the computer program further comprise: updating said current student model with said candidate student model and updating current weights with said new weights in response to said candidate student model generating a better prediction of said observed target than said current student model.
21. The system as recited in claim 20, wherein the program instructions of the computer program further comprise: generating new state features by said updated student model using said new weights; and generating a second set of new weights by said teacher model using said new state features.
22. The system as recited in claim 21 , wherein the program instructions of the computer program further comprise: training said candidate student model using said training data and said second set of new weights; and generating a reward by comparing said updated student model with said candidate student model using said training data and said testing data to determine which is better at predicting said observed target.
23. The system as recited in claim 19, wherein the program instructions of the computer program further comprise: generating a second set of new weights by said updated teacher model using said state features in response to said candidate student model not generating a better prediction of said observed target than said current student model.
24. The system as recited in claim 23, wherein the program instructions of the computer program further comprise: training said candidate student model using said training data and said second set of new weights; and generating a reward by comparing said current student model with said candidate student model using said training data and said testing data to determine which is better at predicting said observed target.
25. The system as recited in claim 18, wherein said stopping rule comprises one or more of the following: reaching a specified number of trials, reaching a specified training time, converging of a prediction accuracy, and a user-initiated termination.
GB2207662.4A 2019-11-14 2020-11-10 Identifying optimal weights to improve prediction accuracy in machine learning techniques Withdrawn GB2603445A (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US16/684,396 US11443235B2 (en) 2019-11-14 2019-11-14 Identifying optimal weights to improve prediction accuracy in machine learning techniques
PCT/IB2020/060581 WO2021094923A1 (en) 2019-11-14 2020-11-10 Identifying optimal weights to improve prediction accuracy in machine learning techniques

Publications (2)

Publication Number Publication Date
GB202207662D0 GB202207662D0 (en) 2022-07-06
GB2603445A true GB2603445A (en) 2022-08-03

Family

ID=75908006

Family Applications (1)

Application Number Title Priority Date Filing Date
GB2207662.4A Withdrawn GB2603445A (en) 2019-11-14 2020-11-10 Identifying optimal weights to improve prediction accuracy in machine learning techniques

Country Status (8)

Country Link
US (2) US11443235B2 (en)
JP (1) JP7471408B2 (en)
KR (1) KR20220066163A (en)
CN (1) CN114616577A (en)
AU (1) AU2020385049B2 (en)
DE (1) DE112020005610T5 (en)
GB (1) GB2603445A (en)
WO (1) WO2021094923A1 (en)

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20210182660A1 (en) * 2019-12-16 2021-06-17 Soundhound, Inc. Distributed training of neural network models
US11551083B2 (en) 2019-12-17 2023-01-10 Soundhound, Inc. Neural network training from private data
WO2021195095A1 (en) * 2020-03-23 2021-09-30 Google Llc Neural architecture search with weight sharing
US20230214950A1 (en) * 2020-04-23 2023-07-06 Nec Corporation Prediction device, prediction method, and recording medium
US11663522B2 (en) * 2020-04-27 2023-05-30 Microsoft Technology Licensing, Llc Training reinforcement machine learning systems
US11620576B1 (en) * 2020-06-22 2023-04-04 Amazon Technologies, Inc. Systems and methods for knowledge transfer in machine learning
US11497001B2 (en) * 2020-11-19 2022-11-08 Kabushiki Kaisha Toshiba Edge-intelligence for stability guaranteed real-time control systems
US20230102489A1 (en) * 2021-09-07 2023-03-30 Samsung Electronics Co., Ltd. Method of load forecasting via knowledge distillation, and an apparatus for the same
US20230196067A1 (en) * 2021-12-17 2023-06-22 Lemon Inc. Optimal knowledge distillation scheme
US11386368B1 (en) * 2022-03-04 2022-07-12 John Schneider Method for matching students with teachers to achieve optimal student outcomes
CN115359062B (en) * 2022-10-24 2023-01-24 浙江华是科技股份有限公司 Method and system for dividing and calibrating monitoring target through semi-supervised example
CN116629133A (en) * 2023-05-30 2023-08-22 山东大卫国际建筑设计有限公司 Prediction method, equipment and medium for building multi-region temperature
CN116564292B (en) * 2023-06-27 2024-02-13 镁佳(北京)科技有限公司 Voice classification model training method, voice classification method, device and equipment
CN116594349B (en) * 2023-07-18 2023-10-03 中科航迈数控软件(深圳)有限公司 Machine tool prediction method, machine tool prediction device, terminal equipment and computer readable storage medium

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170083829A1 (en) * 2015-09-18 2017-03-23 Samsung Electronics Co., Ltd. Model training method and apparatus, and data recognizing method
US20170132528A1 (en) * 2015-11-06 2017-05-11 Microsoft Technology Licensing, Llc Joint model training
US20170372201A1 (en) * 2016-06-22 2017-12-28 Massachusetts Institute Of Technology Secure Training of Multi-Party Deep Neural Network
CN110175628A (en) * 2019-04-25 2019-08-27 北京大学 A kind of compression algorithm based on automatic search with the neural networks pruning of knowledge distillation

Family Cites Families (26)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2009103025A2 (en) * 2008-02-15 2009-08-20 Biometallix, Llc Reactor for extracting metals from metal sulfide containing materials and methods of use
US10318882B2 (en) * 2014-09-11 2019-06-11 Amazon Technologies, Inc. Optimized training of linear machine learning models
US20160071017A1 (en) 2014-10-15 2016-03-10 Brighterion, Inc. Method of operating artificial intelligence machines to improve predictive model training and performance
EP3497625A1 (en) * 2016-08-11 2019-06-19 Twitter, Inc. Aggregate features for machine learning
JP6938980B2 (en) 2017-03-14 2021-09-22 富士フイルムビジネスイノベーション株式会社 Information processing equipment, information processing methods and programs
US11748684B2 (en) * 2017-03-31 2023-09-05 Raytheon Technologies Corp. Predictive analytics systems and methods
JP6820815B2 (en) 2017-09-07 2021-01-27 株式会社日立製作所 Learning control system and learning control method
US10257072B1 (en) 2017-09-28 2019-04-09 Cisco Technology, Inc. Weight initialization for random neural network reinforcement learning
US20190102693A1 (en) * 2017-09-29 2019-04-04 Facebook, Inc. Optimizing parameters for machine learning models
US20190102709A1 (en) * 2017-10-03 2019-04-04 Invight, Inc. Systems and methods for coordinating venue systems and messaging control
WO2019096754A1 (en) 2017-11-20 2019-05-23 Koninklijke Philips N.V. Training first and second neural network models
US10643602B2 (en) 2018-03-16 2020-05-05 Microsoft Technology Licensing, Llc Adversarial teacher-student learning for unsupervised domain adaptation
US11423501B2 (en) * 2018-10-30 2022-08-23 Oracle International Corporation Machine learning for optimal student guidance
US20200134445A1 (en) * 2018-10-31 2020-04-30 Advanced Micro Devices, Inc. Architecture for deep q learning
US11656620B2 (en) * 2018-12-31 2023-05-23 Luminar, Llc Generating environmental parameters based on sensor data using machine learning
US20200218940A1 (en) * 2019-01-08 2020-07-09 International Business Machines Corporation Creating and managing machine learning models in a shared network environment
US11119250B2 (en) * 2019-01-15 2021-09-14 International Business Machines Corporation Dynamic adaption of vessel trajectory using machine learning models
US20200257980A1 (en) * 2019-02-08 2020-08-13 International Business Machines Corporation Training optimization for neural networks with batch norm layers
US11093862B2 (en) * 2019-03-21 2021-08-17 International Business Machines Corporation Locality aware data loading for machine learning
US11586930B2 (en) * 2019-04-16 2023-02-21 Microsoft Technology Licensing, Llc Conditional teacher-student learning for model training
CN110472494A (en) * 2019-06-21 2019-11-19 深圳壹账通智能科技有限公司 Face feature extracts model training method, facial feature extraction method, device, equipment and storage medium
US20200401949A1 (en) * 2019-06-24 2020-12-24 Microsoft Technology Licensing, Llc Optimizing machine learned models based on dwell time of networked-transmitted content items
US11276023B1 (en) * 2019-09-06 2022-03-15 Amazon Technologies, Inc. Machine learning optimization for fraud detection
US11640528B2 (en) * 2019-10-22 2023-05-02 Baidu Usa Llc Method, electronic device and computer readable medium for information processing for accelerating neural network training
US11755946B2 (en) * 2019-11-08 2023-09-12 International Business Machines Corporation Cumulative reward predictor training
US20220366678A1 (en) * 2019-11-13 2022-11-17 Nec Corporation Learning apparatus, learning method, and recording medium

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170083829A1 (en) * 2015-09-18 2017-03-23 Samsung Electronics Co., Ltd. Model training method and apparatus, and data recognizing method
US20170132528A1 (en) * 2015-11-06 2017-05-11 Microsoft Technology Licensing, Llc Joint model training
US20170372201A1 (en) * 2016-06-22 2017-12-28 Massachusetts Institute Of Technology Secure Training of Multi-Party Deep Neural Network
CN110175628A (en) * 2019-04-25 2019-08-27 北京大学 A kind of compression algorithm based on automatic search with the neural networks pruning of knowledge distillation

Also Published As

Publication number Publication date
JP2023501257A (en) 2023-01-18
DE112020005610T5 (en) 2022-09-01
US20210150407A1 (en) 2021-05-20
US11443235B2 (en) 2022-09-13
JP7471408B2 (en) 2024-04-19
US20220292401A1 (en) 2022-09-15
WO2021094923A1 (en) 2021-05-20
AU2020385049B2 (en) 2023-02-02
KR20220066163A (en) 2022-05-23
GB202207662D0 (en) 2022-07-06
AU2020385049A1 (en) 2022-04-28
CN114616577A (en) 2022-06-10

Similar Documents

Publication Publication Date Title
GB2603445A (en) Identifying optimal weights to improve prediction accuracy in machine learning techniques
KR102242516B1 (en) Train machine learning models on multiple machine learning tasks
CN109902672B (en) Image labeling method and device, storage medium and computer equipment
Kaplan et al. Setting expected timelines of fished population recovery for the adaptive management of a marine protected area network
GB2601663A (en) Automated neural network generation using fitness estimation
US20160357790A1 (en) Resolving and merging duplicate records using machine learning
US20140279739A1 (en) Resolving and merging duplicate records using machine learning
US20170213150A1 (en) Reinforcement learning using a partitioned input state space
US20160086498A1 (en) Recommending a Set of Learning Activities Based on Dynamic Learning Goal Adaptation
WO2018201151A1 (en) Neural network optimizer search
CN105046366B (en) model training method and device
US20220357929A1 (en) Artificial intelligence infused estimation and what-if analysis system
Jani et al. A framework of software requirements quality analysis system using case-based reasoning and Neural Network
EP2399202A1 (en) Method and system for calculating value of website visitor
US20200184383A1 (en) User intent classification using a multi-agent reinforcement learning framework
GB2607738A (en) Data augmented training of reinforcement learning software agent
KR20110062200A (en) Apparatus for providing learning contents and method thereof
CA3046475A1 (en) System and method for processing natural language statements
Custode et al. A co-evolutionary approach to interpretable reinforcement learning in environments with continuous action spaces
Garcia et al. Bio-economic management strategy evaluation of deepwater stocks using the FLBEIA model⋆
CN111124916B (en) Model training method based on motion semantic vector and electronic equipment
WO2024068784A1 (en) Reward-model based reinforcement learning for performing reasoning tasks
KR20240034804A (en) Evaluating output sequences using an autoregressive language model neural network
US12033230B2 (en) Development of geo-spatial physical models using historical lineage data
Del Val et al. A team formation tool for educational environments

Legal Events

Date Code Title Description
WAP Application withdrawn, taken to be withdrawn or refused ** after publication under section 16(1)